Gardens and cooking classes in Kon Tum, Vietnam
Here we model an intervention to address malnutrition among ethnic minority groups in Kon Tum, Vietnam. The model follows observations and discussions and some follow up work by partners of the NIFAM (Nutrition Intervention Forecasting and Monitoring) project in 2022 and 2023. A test run of the intervention will be carried out by the Social Policy Ecology research Institute (SPERI) in Kon Tum with smallholder farmers of the Ca Dong ethnic minority group.
Kon Tum garden and cooking class simulation
We use the mcSimulation function to perform a Monte
Carlo simulations to estimate model outputs based on provided parameters
and a model function. The Monte Carlo simulation generates a set of
estimated model outputs based on random input samples, providing a
distribution of potential outcomes.
# Source our model
source("KonTum_Garden_Model.R")## The legacy packages maptools, rgdal, and rgeos, underpinning the sp package,
## which was just loaded, will retire in October 2023.
## Please refer to R-spatial evolution reports for details, especially
## https://r-spatial.org/r/2023/05/15/evolution4.html.
## It may be desirable to make the sf package available;
## package maintainers should consider adding sf to Suggests:.
## The sp package is now running under evolution status 2
## (status 2 uses the sf package in place of rgdal)##
## Attaching package: 'decisionSupport'## The following objects are masked _by_ '.GlobalEnv':
##
## chance_event, discount, vv# Ensure consistent results with the random number generator
# not for each 'run' of the MC simulation but for
# consistency each time we call on the simulation
set.seed(1234)
garden_simulation_results <- mcSimulation(
estimate = estimate_read_csv("inputs_kontum_garden.csv"),
model_function = kontum_garden_function,
numberOfModelRuns = 1000, #run 1000 times
functionSyntax = "plainNames"
)The way we present NPV values can influence decision makers. The same information presented in different ways can even lead to different decisions. Here we derive a plot that compares the decision options as pure NPV values.
source("functions/plot_distributions.R")## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.2 ✔ readr 2.1.4
## ✔ forcats 1.0.0 ✔ stringr 1.5.0
## ✔ ggplot2 3.4.2 ✔ tibble 3.2.1
## ✔ lubridate 1.9.2 ✔ tidyr 1.3.0
## ✔ purrr 1.0.1
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorsplot_distributions(mcSimulation_object = garden_simulation_results,
vars = c("NPV_garden", "NPV_no_garden"),
method = 'hist_simple_overlay',
base_size = 7,
x_axis_name = "Comparative NPV outcomes")
Framing the outcomes
Under Prospect Theory the way we present NPV values can influence decision makers - the same information presented in different ways can lead to different decisions. For example, framing a projected NPV gain as a “reduction in potential loss” might make it more attractive to decision makers due to loss aversion.
Here we plot the distribution for the decision and frame the
projected NPV gain for the ‘decision’ (distributions for the two options
with the NPV values of the no garden option subtracted from
those for the garden). If we display this as a “Reduction
in potential loss” it might be more attractive to decision makers due to
loss aversion, i.e. the party might put more emphasis on avoiding
potential losses than on seeking gains. We can frame our results as a
strategy that minimizes losses rather than one that maximizes gains.
plot_distributions(mcSimulation_object = garden_simulation_results,
vars = "decision",
method = 'hist_simple_overlay',
base_size = 7,
x_axis_name = "Reduction in potential loss")
Summary of results for the decision
Summary
Use gt_plt_summary() from {gtExtras}
library(gtExtras)
library(svglite)
# Subset the outputs from the mcSimulation function (y) to summarize only on the variables that we want.
mcSimulation_summary <- data.frame(garden_simulation_results$x[2:38],
garden_simulation_results$y[1:3])
gt_plt_summary(mcSimulation_summary) | mcSimulation_summary | ||||||
| 1000 rows x 40 cols | ||||||
| Column | Plot Overview | Missing | Mean | Median | SD | |
|---|---|---|---|---|---|---|
| discount_rate | 0.0% | 6.5 | 6.5 | 0.9 | ||
| size_of_garden | 0.0% | 75.2 | 75.2 | 14.9 | ||
| CV_value | 0.0% | 0.3 | 0.3 | 0.1 | ||
| inflation_rate | 0.0% | 7.5 | 7.5 | 1.5 | ||
| if_community_likes | 0.0% | 0.5 | 0.5 | 0.2 | ||
| if_effective_manage | 0.0% | 0.6 | 0.6 | 0.1 | ||
| if_garden_yield_enough | 0.0% | 0.4 | 0.4 | 0.1 | ||
| if_garden_healthy | 0.0% | 0.7 | 0.7 | 0.1 | ||
| if_effective_training | 0.0% | 0.5 | 0.5 | 0.2 | ||
| if_offer_green_space | 0.0% | 0.7 | 0.7 | 0.1 | ||
| if_reduce_polution | 0.0% | 0.4 | 0.4 | 0.1 | ||
| if_biophysical_good | 0.0% | 0.4 | 0.4 | 0.1 | ||
| equipment_cost | 0.0% | 74.9 | 75.0 | 15.2 | ||
| construction_cost | 0.0% | 22.4 | 22.5 | 4.6 | ||
| garden_designing_costs | 0.0% | 12.5 | 12.4 | 1.5 | ||
| compost_starting | 0.0% | 7.4 | 7.4 | 1.5 | ||
| worm_starting | 0.0% | 3.5 | 3.5 | 0.9 | ||
| livestock_costs | 0.0% | 3.5 | 3.5 | 0.9 | ||
| maintaining_labor | 0.0% | 32.5 | 32.4 | 4.7 | ||
| seed_costs | 0.0% | 1.5 | 1.5 | 0.3 | ||
| fertilizer | 0.0% | 1.5 | 1.5 | 0.3 | ||
| plant_protection | 0.0% | 3.5 | 3.5 | 0.9 | ||
| child_veg_access | 0.0% | 7.5 | 7.5 | 1.5 | ||
| child_healthier_choices | 0.0% | 290.6 | 285.5 | 119.1 | ||
| women_veg_access | 0.0% | 7.5 | 7.5 | 1.6 | ||
| women_healthier_choices | 0.0% | 297.1 | 298.6 | 121.3 | ||
| elderly_veg_access | 0.0% | 7.5 | 7.5 | 1.5 | ||
| elderly_healthier_choices | 0.0% | 299.2 | 292.0 | 119.3 | ||
| others_veg_access | 0.0% | 7.5 | 7.5 | 1.6 | ||
| others_healthier_choices | 0.0% | 303.3 | 301.0 | 119.7 | ||
| green_space_value | 0.0% | 149.9 | 149.1 | 30.8 | ||
| reduce_polution_value | 0.0% | 15.2 | 15.2 | 3.0 | ||
| school_event_value | 0.0% | 29.6 | 29.5 | 12.5 | ||
| if_not_fallow | 0.0% | 0.4 | 0.4 | 0.1 | ||
| value_of_non_garden_land_use | 0.0% | 353.4 | 358.2 | 89.5 | ||
| costs_of_non_garden_land_use | 0.0% | 3.0 | 3.0 | 1.2 | ||
| value_of_fallow_non_garden_land | 0.0% | 34.5 | 34.2 | 8.8 | ||
| NPV_garden | 0.0% | 2,744.8 | 2,695.8 | 777.5 | ||
| NPV_no_garden | 0.0% | 751.9 | 197.1 | 826.5 | ||
| decision | 0.0% | 1,992.9 | 2,098.4 | 1,104.6 | ||
Summary of the savings
summary(garden_simulation_results$y$decision)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -1145 1303 2098 1993 2733 5043Summary of costs
summary(garden_simulation_results$y$total_costs)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 232.2 322.1 343.5 343.8 366.5 447.1First year
summary(garden_simulation_results$y$Cashflow_garden1)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 65.86 338.74 436.00 444.09 539.97 908.71Cashflow of the garden option
source("functions/plot_cashflow.R")
plot_cashflow(mcSimulation_object = garden_simulation_results,
cashflow_var_name = "Cashflow_garden")
EVPI
# Subset the outputs from the mcSimulation function (y) by selecting the correct variables be sure to run the multi_EVPI only on the variables that the we want.
mcSimulation_table <- data.frame(garden_simulation_results$x,
garden_simulation_results$y[1:3])
source("functions/multi_EVPI.R")
evpi <- multi_EVPI(mc = mcSimulation_table, first_out_var = "NPV_garden")## [1] "Processing 3 output variables. This can take some time."
## [1] "Output variable 1 (NPV_garden) completed."
## [1] "Output variable 2 (NPV_no_garden) completed."
## [1] "Output variable 3 (decision) completed."source("functions/plot_evpi.R")
plot_evpi(evpi, decision_vars = "decision")## Warning: There are no variables with a positive EVPI. You probably do not need
## a plot for that.Projection to Latent Structure
Here we run a Projection to Latent Structures (PLS) model, a flexible type of regression model. The results of this model can give us some sense of the correlation strength and direction for model variables and our outcome variables.
source("functions/pls_model.R")
pls_result <- pls_model(object = garden_simulation_results,
resultName = names(garden_simulation_results$y)[1],
ncomp = 1)
input_table <- read.csv("inputs_kontum_garden.csv")
source("functions/plot_pls.R")
plot_pls(pls_result, input_table = input_table, threshold = 0)
The full repository can be accessed with the following QR code.
